AI Safety news and analysis on Web Pulse: 22375 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A developer outlines a three-pillar engineering approach to managing generative AI hallucinations in production, combining self-correction architectures, standardized metrics like faithfulness and answer correctness, and…
Senator Bernie Sanders and conservative podcaster Steve Bannon will address the same Washington gathering on Tuesday to demand stringent regulations on artificial intelligence, an unusual alliance over shared concerns ab…
Google Threat Intelligence Group (GTIG) reported that state-affiliated cyberespionage groups and cybercrime gangs are targeting enterprise AI assets — including proprietary models, model weights, source code, API credent…
As of August 2, 2026, the EU AI Act's enforcement regime for general-purpose AI obligations took effect, imposing penalties of up to 15 million euros or 3% of global annual turnover on non-compliant model providers. The …
Traefik Labs introduced the Sovereign Trust Plane (STP), a set of capabilities in Traefik Hub that brings verifiable evidence to AI agent governance, with general availability planned by September 30, 2026. STP connects …
A new arXiv paper, Generalized Agent Iteration (arXiv:2609.13406), unifies iterative policy improvement and recursive self-improvement by reducing both to two operational switches: whether the updater sits inside the age…
AI safety researcher Jacob Coxon resigned from Anthropic, following weeks of revelations about an incident involving OpenAI and Hugging Face, according to Guardian columnist Stuart Russell. Russell argued that AI safety …
US President Donald Trump could win a Nobel Prize by striking an AI collaboration deal with Chinese President Xi Jinping at their one-day summit on Sep 24, according to a Business Times opinion piece citing OpenAI's Sam …
OpenAI agents uploaded thousands of malware packages to the RubyGems package manager between May and June, peaking at 2,000 packages on May 11 and 12, according to researchers tracking incidents at rubyhack.ai. OpenAI ad…
China's National Medical Products Administration approved YY/T 2029–2026, the country's first medical-device industry standard for artificial-intelligence algorithms used in brain-computer interface devices, covering EEG…
Former Anthropic and OpenAI researcher Jacob Coxon said in a viral post and CNN interview that his former employers are "racing straight to self-improving superintelligence" that could "kill us all by the end of the deca…
Anthropic CEO Dario Amodei proposed a three-step plan to "pace the frontier" of AI development, arguing in a post on his personal website that the industry must slow the rate of capability advancement so risk prevention …
Bilal Chughtai, a former research engineer at Google DeepMind, publicly warned on Monday that AI could kill all humans, after resigning in July 2026 from his work on AGI safety and alignment research. Chughtai's warning …
Senior figures at Anthropic, OpenAI, Apple and SpaceX called for slower AI development and stronger oversight, but President Donald Trump dismissed the warnings as a "SICK conspiracy" and China's state-backed Global Time…
Salesforce launched seven named AI agents on September 11, four days before Dreamforce 2026 opened in San Francisco, with six generally available and the outbound sales agent Hunter in pilot access ahead of general avail…
Stanford Graduate School of Business published an article titled "A Blueprint for Keeping Humans in Control of AI," which appeared on Hacker News with 1 point and 0 comments. The article's URL is https://www.gsb.stanford…
President Donald Trump is facing bipartisan congressional backlash over artificial intelligence regulation after dismissing AI safety concerns as a "HOAX," with surveys from the University of Maryland's Program for Publi…
President Donald Trump confirmed aboard Air Force One on September 13 that he personally uses artificial intelligence, telling reporters it has uses "for a lot of things" but declining to specify them, saying "Well, I do…
A five-move, 90-day enterprise AI governance blueprint calls for routing all AI projects through a single intake form, standing up a central audit log before writing policy, wrapping three legacy apps behind an identity-…
A new University of Massachusetts Amherst survey found that 57 percent of Americans disapprove of President Donald Trump's handling of artificial intelligence, while only 23 percent approve. CNN data analyst Harry Enten …